百望股份创始人、董事长、CEO陈杰:AI缺的不只是算力,还有“真值”|2026商业新愿景

Core Insights - The market's attitude towards artificial intelligence (AI) has experienced significant fluctuations, moving from initial enthusiasm to a more cautious approach as businesses recognize the limitations of AI in making critical commercial decisions [3][4]. Group 1: AI Limitations and Business Needs - Current AI models primarily rely on publicly available internet data, which often contains unreliable information, leading to "hallucinations" where AI generates inaccurate responses [5]. - Businesses require concrete data for decision-making, such as repayment ability for loans and fulfillment records for procurement, rather than probabilistic predictions [5][6]. Group 2: Company Strategy and Business Model Shift - The company has shifted its business model from selling software (SaaS) to selling results (RaaS) to meet the evolving needs of enterprises that prefer definitive outcomes over analytical insights [9][10]. - A recent collaboration with Ant Group focuses on intelligent risk control, allowing banks to assess the authenticity of transactions and risk levels without extensive manual verification [11][12]. Group 3: Operational Efficiency and Cost Reduction - The company's systems have significantly reduced manual review costs by approximately 30% in financial reimbursement processes for manufacturing enterprises by automating the identification of fraudulent invoices [13]. Group 4: Future Plans and Data Infrastructure - The company aims to strengthen its data infrastructure in collaboration with the government, facilitating the flow of data across various sectors and ensuring data ownership clarity [15]. - The company is also addressing the challenges faced by Chinese enterprises in international markets, particularly in navigating complex tax and compliance requirements through its TaxSwift platform [16][18]. Group 5: Competitive Advantage - Future competition will hinge on the ability to manage high-value, accurate data rather than merely on the size of AI model parameters, positioning the company as a "calibrator" of large models using real data to provide truthful insights [19].